Comparison of Electroencephalography Signal ProcessingTechniques in Brain-Computer Interface Applications

Brain Computer Interface (BCI) systems analyze the Electroencephalography(EEG) signals and translate a person’s intentions into simple commands. Signal processingmethods are essential in such systems and cover: preprocessing, feature extraction, featureselection and classification. Among these processes a model is trained which can predict thelabels of the new data in order to produce a control signal for applying to a BCI system. Inthis study, after filtering and optimizing the epoch window size, training has been done byextracting the suitable feature vector using spatial patterns obtained by the method ofcommon spatial pattern (CSP). After feature extraction, three classifiers namely SupportVector Machine (SVM), Linear Discriminant Analysis (LDA) and Logistic Regression(LogReg) have been trained separately by extracted features and their performance have beenevaluated at the end. The results show the improvement over the previous work.